Graph Based Approaches Used in Association Rule Mining

Abstract

The Thesis aims at in depth study of Graph based approach to find Association Rule Mining and Association Rule Mining without pre-assign weights and its implementation. In the past, several authors have proposed various association rule mining algorithms with and without pre-assigned weights. The Thesis consists of two parts. Part two consists of graph based association rule mining for primitive as well as items which have concept hierarchy. Complete implementation has been done in C++ on various real life datasets. Part one consisted of association rule mining without pre-assigned weights using HITS algorithm.